BUG: nested use of parameters in definition of arrays in numpy.f2py

[2sn]

Nested use of paramaters in definition of dimension caused problmes.
Issue #5877.

For example

  parameter (i=7)
  common buf(nvar*(nvar+1) * (n + 1))

this is a fix as suggested by pearu on github.

[2sn]

gh-5877

[charris]

LGTM, but needs a test. @pearu Look good to you?

[charris]

Also, put Closes #5877 in the commit message for documentation and to automatically close the issue.

[2sn]

Thanks for looking at my pull requests.

My apologies, but I have not done this - changing the commit and also
squashing the commits for #6306 - for github before. Do I create a new
pull request, or what is the procedure?

-Alexander

On 21/09/15 11:41, Charles Harris wrote:

Also, put |Closes #5877| in the commit message for documentation and to
automatically close the issue.

—
Reply to this email directly or view it on GitHub
#6307 (comment).

[charris]

The trick is to get familiar with git rebase -i HEAD~n, where n is the number of commits you want to change. Here n would be 1. Then you can choose to edit commit messages, squash commits, and other such stuff. The command is pretty self explanatory if you run it from the command line. After that, do a force push to the same branch, i.e.,

git push --force origin <your branch name>.

You can also add new commits and just push the branch, no force flag needed.

[pearu]

LGTM as well.

[charris]

To make a test, look at the existing tests in numpy/f2py/tests and try to copy one of them, or add a code snippet that failed before but now works. Feel free to create a new test file if you feel one is warranted.

[charris]

Closing in favor of #6487 . Thanks @2sn .